Abstract
Introduction: The development of next-generation sequencing has made it feasible to interrogate the entire genome or exome (coding genome) in a single experiment. Accordingly, our knowledge of the somatic mutations that cause cancer has increased exponentially in the last years. MPNs and MDS/MPD are chronic myeloid neoplasms characterized by an increased proliferation of one or more hematopoietic cell lineages, and an increased risk of transformation to acute myeloid leukemia (AML). MPNs and MDS/MPDs are heterogenous disorders, both in clinical presentation and in prognosis. We sought to determine the genetic landscape of Ph-negative MPNs and MDS/MPD through next-generation sequencing.
Methods: Paired DNA (sorted CD66b-granulocytes/skin biopsy) from 102 patients with MPNs or MDS/MPD was subjected to whole exome sequencing on a Illumina HiSeq 2000 platform using Agilent SureSelect kit. Diagnosis included primary myelofibrosis (MF; N=42), essential thrombocythemia (ET; N=28), polycythemia vera (PV; N=12), chronic myelomonocytic leukemia (CMML; N=10), systemic mastocytosis (MS; N=6), MDS/MPD-Unclassified (N=2) and post-MPN AML (N=2). Tumor coverage was 150x and germline coverage was 60x. Somatic variants calls were generated by combining the output of Somatic Sniper (Washington University), Mutect (Broad Institute) and Pindel (Washington University). The combined output of these 3 tools was further filtered by in-house criteria in order to reduce false-positive calls (minimum coverage at both tumor/germline ≥8 reads; fraction of reads supporting alternate allele ≥10% in tumor and ≤10% in germline; ratio of allele fraction tumor:germline >2; excluding mutations seen in SNP databases). All JAK2 and CALR mutations were validated through Sanger sequencing. Validation of other somatic mutations is currently underway. Analysis of driver mutations was made with the Intogen web-based software, using the Oncodrive-FM and Oncodrive-cluster algorithms (www.intogen.org). Significantly mutated genes were considered as those with a q-value of <0.10.
Results: We identified a total of 309 somatic mutations in all patients, with each patient having an average of 3 somatic abnormalities, fewer than most solid tumors that have been sequenced so far. Mutations occurred in 166 genes, and 40 of these were recurrently somatically mutated in Ph-negative MPNs. By the Oncodrive-FM algorithm, the following genes were identified as the most significantly mutated driver genes in Ph-negative MPNs and MDS/MPDs (in order of significance): CALR, ASXL1, JAK2, CBL, DNMT3A, U2AF1, TET2, TP53, RUNX1, EZH2, SH2B3 and KIT. By the Oncodrive-cluster algorithm, which considers clustering of mutations at a hotspot, the following genes were significantly mutated: KIT, JAK2, SRSF2 and U2AF1. Somatic mutations were seen in genes that are mutated at a low frequency in Ph-negative MPNs, including ATRX, BCL11A, BCORL1, BIRC5, BRCC3, CSF2RB, CUX1, IRF1, KDM2B, ROS1 and SUZ12.
Consistent with the clinical phenotype, 96 patients (94%) had mutations that lead to increased cellular proliferation, either through activation of the JAK-STAT pathway (e.g. JAK2, CALR) or mutations that activated directly or indirectly signaling by receptor tyrosine kinases (e.g. FLT3, KIT, CBL). Besides biological pathways regulating cell proliferation, the most commonly implicated pathways included regulation of DNA methylation (e.g. DNMT3A, TET2), mRNA splicing (e.g. U2AF1, SRSF2) and histone modifications (e.g. ASXL1, EZH2), seen in 27%, 25% and 22% of patients, respectively. Abnormalities in these 3 pathways were more often seen in MF, MDS/MPD and CMML, as compared to PV and ET (65% vs. 20%; p<0.0001).
Conclusions: Our study represents one of the largest series of patients with these neoplasms evaluated by whole exome sequencing, and together with the published data helps to delineate the genomic landscape of Ph-negative MPNs and MDS/MPDs. The majority of the most frequent mutations seen in Ph-negative MPNs have already been reported. Nevertheless, there are several low frequency mutations that need to be further studied and functionally validated in vitro and in vivo for a deeper knowledge of the pathophysiology of MPNs. Besides activation of cellular proliferation, abnormalities of DNA methylation, histone modification and mRNA splicing emerge as the most important biological pathways in these disorders.
Disclosures
No relevant conflicts of interest to declare.
Whole-Exome Sequencing Identifies Rare and Low-Frequency Coding Variants Associated with LDL Cholesterol
